Hi, as a potential contributor I would like to understand more what this project does and some improvements on how the outcome is shared with the community!

Especially the following:
1. The link to the "about" page, https://boinc.multi-pool.info/latinsquares/about.php, should be fixed and filled with relevant background info. Especially some pointers hinting on what algorithmic approach is used in this project!
2. Is this project related anyhow to finding the 10th entries in the OEIS series A274806 and A274171?
3. How is the result planned to be exposed? Is it an ever growing "database" over found squares? Can this list be found somewhere? We have no hope of storing a complete set judging from the above mentioned 9th entries in the series above or am I missing something? It would at least be nice if each participant could be presented with a number of found squares. Compare with the project Amicable Numbers for instance!
4. I see in some message threads there are files to download from earlier results, but how to parse these files into actual matrices? They contain lines like

1ARFebb6iN7XBRSnJQTzX5S4B
1ARG1NTzWQY7vqHoPChPf5MgS
1AVMSmGkqS4LpQsWMLezvoiey
1AnY4ZgMPJ9hSA8t19LKc3yP
 . . .
 


Here it would be nice with a short python snippet that parses such a file and spits out the squares!

Thanks!

---

hi,

1) the missing page is probably and incompatibility from previous Boinc core (2016) with actual (2025) used into this new server, so the migration from previous to this core has not done it correctly (this is still a todo fix in queue)

2) ODLK1 is the "European" version of ODLK that is Russian based, so it uses the same algorithm (odlk3 and odlkmax) with a set of different key tree to explore different solutions. Description of algorithm is still in ODLK forum and even in old thread onto this forum (however actually the Russian characters are not shown correctly into this server and needs to be fixed in DB).
Solutions are catalogue based onto propriety, like kf1, kf2, ...kf10 (the more upper, the better solution is find)

3)
Solution are released after 1 or 2 years from previous release and the database is a growing one.
for example now we have:

kf1: 4,0GB
kf2: 8,3MB
kf3: 15,1KB
kf4: 15,7KB
kf6: 848B
kf8: 848B
kf10: 212B
Ortogon 8GB

4)
That is a compressed text version based onto encrypt/decrypt programs released in ODLK.

However the new database release up to 2024 (now being processed) will have the square directly accessible from the archive itself. It will be posted as soon as processing ended.

Else new database will separate solution by type kf1, kf2... and no more aggregated as before.

Thanks for the answer! Some further comments,

Description of algorithm is still in ODLK forum and even in old thread onto this forum

Maybe a link to that forum post would be good to have on the front page at least?

However the new database release up to 2024 (now being processed)

Ok, so this will be the first release of ODLK1 results in like 7 years?

More questions for you.

* How do you actually define orthogonal diagonal LS? Searching the net they seem to have to come in pairs or in larger groups where they are all mutually orthogonal. Is this so in your case as well and will that be reflected in the database?

* The 5 matrices on the home page are shown without commenting text. Are those examples from this or another project? How does the orthogonality show up in those?

BRs,

---